Labsco
trayders logo

Trayd

★ 34

from trayders

Trade Robinhood with natural language in Claude Code

🔥🔥🔥✓ VerifiedFreeQuick setup
#email#automation#cloud#devops#git#finance#security

Trayd

Trade Robinhood from Claude. Portfolio analysis, real-time quotes, and order execution through natural language.

MCP Users Orders

Demo

Metrics

User Growth

MCP Calls

What You Can Do

CapabilityExample
Portfolio analysis"What's my portfolio worth?" / "Which positions are up today?"
Real-time quotes"What's NVDA trading at?" — works 24/7
Buy and sell"Buy 10 shares of AAPL" / "Sell my TSLA position"
Limit orders"Place a limit order for TSLA at $400"
Ladder orders"Set 5 ladder buys for NVDA from $180 to $175"
Short selling"Short 10 shares of GME" / "Check short availability for AMC"
Multi-account"List my accounts" / "Show portfolio for account 2"
Batch orders"Buy $500 worth of each: AAPL, GOOGL, MSFT"
Cancel orders"Cancel all my open orders"
Persistent memory"Remember my NVDA thesis" — saved to your private wiki, recalled every session

All limit orders default to 24-hour extended trading. Quotes are served 24/7 — live Robinhood data during market hours, with automatic fallback to a partnered market data provider overnight.

Why Trayd?

Setting 5 ladder limit orders on Robinhood takes 50+ clicks and 5 minutes.

With Trayd: "Set 5 ladder buys for NVDA from $180 to $175" — one sentence, 10 seconds.

What makes Trayd different from other trading MCPs:

  • Works in your browser — the only trading MCP that works on claude.ai (no CLI needed)
  • Remote server — nothing to install, no Python, no Docker, no API keys
  • Full trading — not read-only. Buy, sell, short, batch, ladder
  • Multi-account — manage multiple Robinhood accounts from one connection
  • 24/7 quotes — dual-source pricing even when markets are closed
  • Persistent memory — a private wiki Claude reads and writes, so it remembers your portfolio and strategy across every session

Persistent Memory

Trayd gives Claude a private, persistent knowledge base — your LLM wiki. Tell it something durable ("my cost basis on NVDA is $140", "I only sell covered calls on Fridays", "my thesis on uranium") and Claude saves a markdown note that survives across every future session and conversation — on both claude.ai and Claude Code.

  • It compounds — each conversation starts from everything Claude already knows about your portfolio, strategy, and preferences, instead of from scratch.
  • You own it — notes are private to your account. Claude and you edit the same files, so you stay in control of what it remembers.
  • Examples: "Save my rebalancing rules." / "What's my thesis on SMH?" / "Update my watchlist."

Trading Bot

Claude Code's /loop command turns Trayd into an always-on trading agent:

Copy & paste — that's it
"Check my positions every 5 minutes. If anything drops 3%, sell it."
"Monitor NVDA. If it hits $130, buy 20 shares."
"Rebalance my portfolio to equal weight every morning at 9:35 AM."

No code. No scripts. Just natural language rules that Claude executes on a schedule.

Security

Copy & paste — that's it
You → Claude → Trayd Server → Robinhood API
  • Robinhood credentials pass directly to Robinhood's API — never stored
  • Access tokens held in memory only — wiped on logout, restart, or redeployment
  • OAuth 2.1 with PKCE for MCP authentication
  • Robinhood phone 2FA on every login — you approve each session on your device
  • Infrastructure: AWS ECS Fargate + Cloudflare Tunnel + HTTPS everywhere

Verify it yourself: Place a test order that won't fill — "Limit buy 1 NVDA at $50" — check your Robinhood app, you'll see it. Cancel from either side. Zero risk.

Traction

  • 368 traders signed up, 160+ active every month
  • 2.4M+ MCP calls served — 2.1M+ live quotes and 87K+ orders executed
  • Power users running fully automated strategies via /loop
  • Used on both claude.ai and Claude Code

Disclaimer

This software is provided as-is. You are solely responsible for all trades placed through your account. Not financial advice. Not affiliated with Robinhood Markets, Inc.

Email: team@trayd.ai | Issues: trayders/trayd-mcp